Multiscale reduction improves brain MRI use with 2D foundation models
Reduce, Then Encode: Multiscale Volumetric Reduction for 2D Foundation Models in Brain MRI
Computer Vision and Pattern Recognition
Summary
Brain MRI scans are 3D images, but many advanced image analysis tools are designed for 2D pictures. The authors introduce a way to compress each 3D brain image into a few 2D pieces while keeping important details. These pieces are then processed by existing 2D models without retraining them, making analysis more efficient and effective. This method works well across several different brain MRI datasets and adapts to data from other sources too.
What this means in practice
- •For radiology software developers: Incorporate 3D brain MRI analysis into existing 2D AI models by compressing volumes into informative 2D inputs without retraining encoders.
- •For medical image processing teams: Improve performance and generalization of brain MRI classification tasks by using multiscale reduction before feeding data to 2D pretrained models.
Authors
Dexuan Ding, Yuankai Qi, Bogong Wang, Luping Zhou, Amin Beheshti
Abstract
Pretrained 2D foundation models offer a practical alternative to dedicated 3D pretraining for brain structural magnetic resonance imaging (sMRI), but their use on volumetric data requires bridging the mismatch between a 2D encoder and a 3D volume input. Existing methods typically encode slices independently and integrate their features afterwards. We introduce Multiscale Volumetric Reduction (MVR), a reduce-then-encode approach that compresses each anatomical view from (D) slices into (M << D) complementary 2D components before foundation-model encoding. MVR combines an uncentered-PCA base component derived from the original through-plane intensities with residual detail components constructed from multiscale spatial descriptors. The reduction is estimated from the training volumes without diagnostic labels or gradient-based optimization and remains fixed thereafter. The resulting components are independently processed by a shared frozen 2D foundation model and concatenated for linear probing. Under this frozen-encoder setting, MVR achieves strong overall performance across ADNI, OASIS, and ABIDE relative to the evaluated 2D-to-3D adaptation methods and simple input-reduction baselines, while also generalizing strongly from ADNI to AIBL.